Lysozyme Dispersed Single-Walled Carbon Nanotubes: Interaction

Apr 18, 2012 - Liquid dispersions of 0.5 wt % LSZ–0.1 wt % SWNT were homogenized via vortex mixing, and ∼10 μL were pipetted onto a clean, alumin...
2 downloads 7 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCC

Lysozyme Dispersed Single-Walled Carbon Nanotubes: Interaction and Activity Daniel W. Horn,† Kathryn Tracy,† Christopher J. Easley,‡ and Virginia A. Davis*,† †

Department of Chemical Engineering, Auburn University, 212 Ross Hall, Auburn, Alabama 36849, United States Department of Chemistry & Biochemistry, Auburn University, 179 Chemistry Building, Auburn, Alabama 36849, United States

‡

S Supporting Information *

ABSTRACT: The noncovalent interaction between lysozyme (LSZ) and single-walled carbon nanotubes (SWNT) was probed using a combination of methods including scanning electron microscopy, UV−vis spectroscopy, Raman spectroscopy, circular dichroism (CD), and fluorescence anisotropy. In addition to supporting the previously hypothesized importance of π−π stacking, the results of this research suggest that the key interaction is between the hydrophobic tryptophan residue within LSZ and the sidewall of SWNT. The hydrophobic interaction is so critical to dispersion stabilization that increasing SWNT hydrophilicity through oxidation actually reduces dispersibility in aqueous LSZ. The combination of the strong LSZ−SWNT interaction with the inherent stability of LSZ and preservation of secondary structure enable retention of antibacterial activity both in solution and solid assemblies. This work provides a foundation for advancing understanding and design of materials that combine carbon nanotube properties with natural enzymatic activity.

■

INTRODUCTION Carbon nanotubes (CNT) have been the subject of intense, multidisciplinary research for two decades because of their unique combination of electrical, mechanical, optical, and thermal properties.1 The combination of carbon nanotubes and Nature’s toolbox2 is particularly intriguing because biomolecules such as double-stranded DNA, proteins, and enzymes facilitate CNT dispersion in aqueous solvents and can augment the properties of bulk materials assembled from CNT dispersions.3,4 The most predominantly studied biomolecule in combination with single-walled carbon nanotubes (SWNT) is DNA due to its ability to sort SWNT,5 favorable interactions,6 interfacial properties,7 and phase behavior.8,9 In addition, DNA−SWNT dispersions have been assembled into sensors,7,10−13 films,9 and fibers.14,15 Understanding the interactions between proteins and CNT is important for expanding the biological, medical, and dental applications of CNT materials as well as understanding potential interactions with CNT surfaces in vivo.16−18 Since CNT are difficult to disperse solid materials and proteins are typically studied in solution, there has been a need to develop new protocols to achieve dispersion and characterize interactions.19 There are two approaches to immobilizing proteins on CNT: covalent and noncovalent. Several researchers have used EDC/ NHS chemistry to covalently attach enzymes such as lysozyme (LSZ),20,21 horseradish peroxidase,4 and organophosphate hydrolase22 to CNT. However, covalent attachment of proteins to CNT has the disadvantages of increasing processing time and cost. In addition, covalent functionalization disrupts the intrinsic sp2 bond structure that is responsible for CNT’s outstanding mechanical, electrical, and thermal properties. © 2012 American Chemical Society

Therefore, understanding the noncovalent interactions between proteins and pristine CNT is important for producing materials that combine protein function with the intrinsic properties of pristine CNT. Combinations of LSZ and CNT are of particular interest due to both LSZ’s ability to disperse CNT and its natural activity against Gram-positive bacteria such as Staphylococcus aureus. In addition, both molecular modeling16 and experimental studies have shown that LSZ has the potential to sort nanotubes based on diameter. LSZ has stronger interactions with larger diameter carbon nanotubes, and fractionation of double-walled carbon nanotubes based on diameter has been demonstrated.23 LSZ is a relatively small, globular protein comprising a single polypeptide chain of 129 amino acid residues; it has a molecular weight of approximately 14 000 g mol−1 and a diameter of approximately 3 nm.24,25 LSZ’s antibacterial activity is due to its ability to catalyze the hydrolysis of 1,4-β-glycosidic linkages between N-acetylmuramic acid and N-acetylglucosamine, which are components of the cell wall peptidoglycan of many bacteria.26 LSZ’s four disulfide bridges stabilize the secondary structure and enable it to retain activity over a broad pH and temperature range. In addition, its isoelectric point is 11 making it stable in electrostatic solvents at neutral pH. Its denaturation temperature is 76 °C, and it retains enzymatic activity in both its native and partially denatured states.27 Several researchers have demonstrated the ability to disperse pristine, purified SWNT in aqueous LSZ with the aid of Received: January 8, 2012 Revised: April 15, 2012 Published: April 18, 2012 10341

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348

The Journal of Physical Chemistry C

Article

sonication.17,24,28−30 LSZ−SWNT dispersions have also been used to produce mechanically strong antibacterial films.28 However, while there have been several studies on LSZ− SWNT dispersions, the details of the interaction are not well understood. Nepal and Geckler hypothesized that the size of SWNT facilitated them into fitting into LSZ’s hydrophobic pocket and that the primary interaction was therefore π−π stacking between hydrophobic residues.24 Matsuura et al. echoed this hypothesis and further suggested that partial unfolding during sonication facilitated dispersion,17 and Xie et al. probed LSZ denaturation using circular dichroism (CD) and Raman spectroscopy.30 In the present work, we provide additional evidence of the importance of hydrophobic interactions and π−π stacking. In addition, we explore the hypothesis that the binding mechanism is between tryptophan (aromatic amino acid residue) and the SWNT sidewall. LSZ contains three hydrophobic residues within its core: methionine, phenylalanine, and tryptophan. Of these three amino acids, it is less likely that methionine interacts with the SWNT sidewall since it does not possess any aromatic rings to enable π−π stacking with SWNT. Phenylalanine and tryptophan both possess aromatic rings, but phenylalanine’s benzene group is less likely to interact than tryptophan’s aromatic indole group due to its larger size and higher electrostatic potential.31 The hypothesis that the primary interaction is π−π stacking between the tryptophan and SWNT sidewall is supported by spectroscopic evidence, particularly anisotropic fluorescence measurements. In addition to providing new insights into the LSZ−SWNT interaction, we highlight the excellent retention of LSZ activity in both liquid dispersions and solid materials.

water and thermal gravimetric analysis (see Supporting Information). Scanning Electron Microscopy. A JEOL (Tokyo, Japan) 7000F FE-SEM with energy dispersive X-ray spectroscopy was used for performing scanning electron microscopy (SEM) to qualitatively determine the presence of a molecular interaction between LSZ and SWNT. Liquid dispersions of 0.5 wt % LSZ− 0.1 wt % SWNT were homogenized via vortex mixing, and ∼10 μL were pipetted onto a clean, aluminum SEM sample stub. The sample was then vacuum-dried at 30 mmHg and 80 °C for 24 h. After the sample was dried, it was gold-coated using a PELCO (Redding, CA) SC-6 Sputter Coater to increase the electrical conductance of the surface, which enhances the optical resolution. Atomic Force Microscopy. A Pacific Nanotechnology (Berkley, CA) Nano-R SPM was used for performing atomic force microscopy (AFM) to quantitatively determine the length and diameter of the pristine SWNT used during research. Approximately 5 μL of liquid 0.25 wt % LSZ−0.01 wt % SWNT supernatant dispersion was pipetted onto a cleaned, molecularly flat silicon stub and dried in a vacuum oven at 30 mmHg and 80 °C for 24 h. The AFM was operated in a static, noncontact mode. Each scan was of a 5 μm2 area at a resolution of 256 and a scan speed of 0.5 Hz. Spectroscopy. A Cary (Santa Clara, CA) 3E UV−vis Spectrophotometer was used in this work. Scans were performed at room temperature at wavelengths from 200 to 800 nm at a scan rate of 600 nm/s and a resolution of 1 nm. Absorption spectroscopy was used to characterize the LSZ solutions and LSZ−SWNT dispersions and quantitatively determine any shifts in absorption wavelength or intensity. Liquid dispersions of 0.5 wt % LSZ−0.1 wt % SWNT were diluted at a ratio of 100 μL of solution/dispersion to 5 mL of DI water and were then homogenized via vortex mixing. Approximately 300 μL of LSZ solution or LSZ−SWNT dispersion samples were then pipetted into 1 mm path length quartz UV−vis cell for scanning. A Renishaw (Gloucesterchire, UK) in Via microRaman was used for Raman spectroscopy. Samples were prepared by drying in a vacuum for 24 h. The samples were then placed on a glass slide by suspending on double-sided tape. Prior to running tests on all samples, each laser was calibrated using a silicon reference. Each sample was scanned by both a 514 and 785 nm laser to probe both the semiconducting and metallic CNT present. Each run by each laser consisted of 10 scans over a Raman shift from 100 cm−1 to 3200 cm−1. A TA Instruments (New Castle, DE) Q50 Thermogravimetric Analyzer (TGA) coupled to a Nicolet (Waltham, MA) iS10 Fourier-Transform-Infrared spectrometer (FTIR) was used for determination of component concentration in solid assemblies, evidence of oxidation of oxidized SWNT, and presence of interaction between LSZ and SWNT. ATR-FTIR was performed on dried samples of SWNT, LSZ, LSZ−SWNT mixture, and LSZ−SWNT supernatant on a germanium plate. Each sample run consisted of 64 scans from 500 cm−1 to 3700 cm−1. A Perkin-Elmer (Waltham, MA) LS 55 Luminescence Spectrometer was used for fluorescence measurements. Emission spectra were taken at room temperature scanning from 300 to 500 nm with a 0.5 nm resolution and a scan rate of 500 nm/min with an incident angle of 90°. The excitation wavelength was set to 280 nm to probe the quenching of the tryptophan within LSZ when coupled with SWNT. Liquid

■

EXPERIMENTAL METHODS Materials. The SWNT used in this research were Unidym (Sunnyvale, CA) 187.1 and 187.2 SWNT made via the high pressure carbon monoxide (HiPco) process. The HiPco SWNT had residual catalyst of ∼1% based on thermogravimetric analysis. The average diameter of the HiPco SWNT was ∼1.24 nm, and the average length was ∼1042 nm as determined by atomic force microscopy. Additional nanotube characterization data is in the Supporting Information. Dialyzed, lyophilized LSZ from chicken egg white was purchased from Sigma Aldrich (St. Louis, MO) and used as received. The LSZ was determined to be 90% pure via spectroscopy with the impurities being buffer salts of sodium acetate and sodium chloride. Sample Preparation. LSZ solutions were prepared by adding 500 mg of LSZ to 100 mL of deionized water and mechanically stirring using a magnetic stir plate and 1 in. Teflon stir bar set at ∼200 rpm. LSZ−SWNT dispersions were used as either bulk mixtures or supernatants. LSZ−SWNT bulk mixtures were prepared by adding 100 mg of SWNT to the 0.5 wt % LSZ solution and pulse sonicating in an ice bath for 30 min with pulses of 5 s on and 2 s off. After sonication, the bulk mixture preparation was complete. These initial dispersions of 0.5 wt % and 0.1 wt % SWNT were centrifuged for 3 h at 17 000g to obtain a supernatant of 0.220 wt % LSZ and 0.015 wt % SWNT. Oxidized SWNT were produced by sulfuric/nitric acid oxidation. The desired amount of pristine SWNT was added to an Erlenmeyer flask and a mixture of 3:1 sulfuric acid−nitric acid was added at a concentration of 1.5 mL total acid per 1 mg of pristine SWNT. The acid−SWNT solution was then bath sonicated for 1 h. After sonication, the tubes were neutralized through hexane−water extraction prior to being filtered. Oxidation was evidenced by dispersibility in 10342

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348

The Journal of Physical Chemistry C

Article

solutions of 0.5 wt % LSZ and dispersions of 0.5 wt % LSZ−0.1 wt % SWNT were diluted at a ratio of 50 μL of solution/ dispersion to 5 mL of DI water. The diluted solutions were then homogenized via vortex mixing. The cuvette used for the emission fluorescence testing was a 4-sided polycarbonate cuvette with a 1 cm path length. Anisotropic fluorescence data was collected at an excitation wavelength of 295 nm with an emission wavelength of 345 nm. The excitation and emission slits were each set at 15 nm where a G-factor was calculated at 1.581. A total of 33 anisotropy measurements were performed in a 400 μL cylindrical cuvette with a 5 mm path length. A JASCO (Easton, MD) J-810 Spectropolarimeter was used for circular dichroism (CD) measurements. Scans were taken at room temperature with a working spectral window of 190 to 400 nm with a scan rate of 100 nm/min and a resolution of 1 nm. CD was used to determine the specific interaction between LSZ and SWNT through the effect that SWNT has on the secondary and tertiary structure of LSZ. Liquid dispersions of 0.5 wt % LSZ−0.1 wt % SWNT were homogenized via vortex mixing. Because of dispersion turbidity, approximately 5 μL of LSZ−SWNT dispersion were then pipetted into a 0.5 mm path length quartz CD cell for scanning. Turbidmetric Analysis of Antibacterial Activity. A Cary (United States) 3E UV−vis Spectrophotometer was used for the turbidimetric analysis. First, a 66 mM solution of potassium phosphate buffer was adjusted to pH 6.24 with potassium hydroxide at room temperature. Next, a bacterial cell suspension in the potassium phosphate buffer containing 0.015 wt % Micrococcus lysodeikticus (Sigma Aldrich) was prepared. Next, 1 cm path length quartz cuvettes were filled with 2.5 mL of bacterial suspension, at which point 0.1 mL of buffer solution or liquid dispersions are added. Finally, UV−vis was performed on the buffered bacterial suspension followed by UV−vis of a buffered bacterial suspension containing the liquid dispersions. The kinetic scans were at 450 nm for 5 min. This turbidimetric analysis was used to determine the lytic activity of all dispersions and macroscopic assemblies.

Figure 1. Scanning electron micrograph of a dried 0.5 wt.% LSZ−0.1 wt % SWNT aqueous supernatant illustrating the direct interaction between SWNT (opaque) and LSZ (translucent) along with swelling of the biopolymer LSZ layer; (a) 50 000×, (b) 200 000×, (c) 300 000×, and (d) 400 000×.

cannot be determined from this image due to the limits of resolution, but previous studies have shown that dispersions prepared by the same method dispersed SWNT as individuals and small bundles.28 Since drying can affect both morphology and the specificity of interaction, additional methods were used to confirm the nature of the SWNT−LSZ interaction. Immobilization of LSZ on SWNT was also evidenced by attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR) and Raman spectroscopy. Figure 2

■

RESULTS AND DISCUSSION The close association between LSZ and SWNT was visualized using scanning electron microscopy (SEM). Although SEM is primarily intended for scanning the surface of molecules, a certain amount of transmission occurs. The ability of electrons to scan a surface versus transmit through the surface is dependent upon the electronic conductivity of the molecules being imaged. If the material is highly conductive, then the electrons will scan the surface giving a high resolution, opaque image. If the material is not conductive, the electrons can transmit through the material giving a low resolution, partially translucent image. Since SWNT are significantly more electrically conductive than LSZ, SWNT appear opaque under SEM, while LSZ appears translucent. As shown in Figure 1, small SWNT bundles were intimately surrounded by LSZ; no other materials were present in the sample. In addition, as the electrical energy from the electrons was transferred to the LSZ in the form of thermal energy, the LSZ began to swell. This is a common effect in SEM of proteins and biological polymers. Figure 1a was taken last sequentially, and it can be seen by comparison with Figure 1b,c that the empty spaces between the coated SWNT have been filled in by the swelling of the bound LSZ. Further evidence of this effect is shown by the size ratio of LSZ to SWNT increasing by approximately 3-fold from Figure 1b to d. The exact bundle size

Figure 2. ATR-FTIR spectrum of SWNT, LSZ, 0.5 wt.% LSZ−0.1 wt. % SWNT mixture, and the resulting supernatant. The LSZ and LSZ− SWNT dispersions show the amide I and II peaks as well as the characteristic amino radical peak characteristic of LSZ confirming adsorption of LSZ to SWNT.

shows the ATR-FTIR spectra of dried samples of SWNT, LSZ, the aqueous LSZ−SWNT mixture prior to centrifugation, and the resulting LSZ−SWNT supernatant. The LSZ, LSZ−SWNT mixture, and resulting supernatant spectra exhibited the amide I (∼1650 cm−1) and amide II (∼1535 cm−1) peaks characteristic of proteins and enzymes. In addition, the LSZ−SWNT mixture and supernatant overlay the LSZ peak at ∼3300 cm−1, which is 10343

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348

The Journal of Physical Chemistry C

Article

Figure 3. Raman spectra with (a) 514 nm and (b) 785 nm excitation of SWNT, the LSZ−SWNT mixture, and the resulting supernatant. The full spectra show the RBM and the D, G, and G′ peaks characteristic of stretching modes within SWNT; the insets highlight the RBM peaks.

Table 1. Peak Positions of RBM, D, G, and G′ for Each Dispersion Tested along with the D:G, G:G′, and D:RBM Ratios from the Raman Spectra of 514 nm Excitation Lasera 514 nm laser −1

Raman shift (cm )

a

intensity ratios

sample

RBM

D

G

G′

D−G

G−G′

D−RBM

SWNT LSZ−SWNT mixture LSZ−SWNT supernatant

263 263 263

1338 1335 1333

1589 1589 1594

2636 2654 2645

0.126 0.136 0.157

1.899 1.374 1.008

0.861 0.657 0.449

The shifts in G and G′ and the decrease of the G:G′ intensity ratio are indicative of charge transfer between the LSZ and SWNT.

3a,b show the RBM peaks. All obvious RBM peaks present for SWNT used in the dispersions are present in both the mixture and supernatant with similar relative intensities. This suggests that, in this work, there was no significant fractionation based on diameter. The D:G intensity ratios varied only slightly between SWNT (0.13), LSZ−SWNT mixture (0.14), and LSZ−SWNT supernatant (0.16). This indicates that the immobilization was through a noncovalent interaction; covalent functionalization would have significantly increased the D:G ratios. The noncovalent interaction is further evidenced by the linearity of the data of D:RBM versus D:G for each excitation laser (see Supporting Information).32 As shown in Figure 3b, the LSZ−SWNT supernatant Raman spectra had significant luminescence interference from 600 cm−1 to 3200 cm−1. Luminescence is associated with individual, noncovalently functionalized SWNT. As the concentration of individual SWNT increases within a dispersion, the luminescence intensity increases dramatically, and the Raman scattering signal will essentially be buried under the luminescence.33−35 Therefore, the Raman spectra of LSZ−SWNT supernatant at 785 nm excitation further validates the noncovalent functionalization of SWNT dispersed as individuals by LSZ. Another interesting facet of the Raman spectra is that several features suggest a charge transfer between the LSZ and SWNT in both the mixture and supernatant.36−39 Both the G38,39 and G′36,37 peaks have been used to elucidate doping effects. While luminescence prevents accurate determination of differences in peak location and intensity using 785 nm excitation, the spectra using 514 nm supports charge transfer (see Table 1). Compared to the pristine SWNT, the mixture and supernatant show an upshift in the G′ peak, from 2636 cm−1 to 2654 and 2645 cm−1, respectively. In contrast, the G:G′ ratio for the

the overtone to the amide I peak and is characteristic of amino radicals present in each sample. Furthermore, the LSZ−SWNT mixture and supernatant show increasing absorbance from ∼1500 cm−1 to ∼700 cm−1 similar to that seen in pristine SWNT. This result clearly indicates that SWNT were stabilized in solution by the presence of LSZ, and thus carried over into the dry sample, in agreement with previous work.17,28−30 The data also shows that a portion of SWNT was lost during centrifugation, but the relative increase between 1200 cm−1 and 700 cm−1 in the supernatant (compared to the pure LSZ) shows that LSZ-stabilized SWNT were still present in the supernatant. Interestingly, the peak (1000 to 900 cm−1) in the SWNT IR absorption band was red-shifted in the dispersed SWNT, where there is a constant increase through 700 cm−1. This is an effect of the aqueous dispersion of LSZ since the dispersed SWNT spectrum is distinctly different from the spectra of noninteracting, dry mixed LSZ and SWNT (see Supporting Information). This ATR-FTIR data thus supports that there is an interaction between LSZ and SWNT. As described below, this interaction was further interrogated using other spectroscopic methods. Figure 3 shows the Raman spectra (514 and 785 nm excitation) of the LSZ−SWNT mixture and supernatant, as well as the initial SWNT material. Several facets of these spectra support the existence of noncovalent functionalization of individual SWNT with LSZ. The spectra show characteristic radial breathing modes (RBM) from 100 cm−1 to 400 cm−1, the D peak due to sp3 hybridized carbons at ∼1335 cm−1, the G peak due to tangential sp2 hybridized carbons at ∼1589 cm−1, and the G′ peak (second overtone of the D peak) at ∼2645 cm−1. It has been previously reported that LSZ preferentially adsorbs to larger diameter nanotubes.23,30 The insets of Figure 10344

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348

The Journal of Physical Chemistry C

Article

Figure 4. (a) UV−vis absorbance spectrum of 0.5 wt % LSZ−0.1 wt % SWNT supernatant showing van Hove singularity peaks characteristic of individually dispersed SWNT and inset showing the characteristic LSZ maximum absorbance at 280 nm due to the tryptophan residue. (b) UV−vis absorbance difference spectra of 0.5 wt % LSZ−0.1 wt % SWNT mixture and 0.5 wt % LSZ−0.1 wt % SWNT supernatant. The absorbance difference intensities have been down shifted to the LSZ solution (zero) beginning at 348 nm to highlight the hyperchromic shift due to the adsorption of LSZ on SWNT.

pH values of 5.3 and 5.6, respectively. For SWNT-Ox, the mixtures and supernatants had pH values of 7.4 and 7.2, respectively. On the basis of Nepal and Geckler, these higher pH values could account for some reduced dispersibility but not the nearly 4-fold reduction. In addition, buffering the SWNT-Ox mixture with 66 mM KH2PO4 to obtain a pH of 5.2, similar to the pristine SWNT, resulted in a supernatant with pH of 5.3 but an even lower SWNT-Ox concentration of 19.0 mg/ L, which is contrary to the expectations based on Nepal and Geckler.24 Therefore, electrostatic effects are not the primary reason for SWNT-Ox having a lower supernatant concentration than pristine SWNT. Most likely, the reduced dispersibility was due to SWNT-Ox’s reduced potential for hydrophobic interaction with LSZ. This indicates that the hydrophobic interaction between LSZ and SWNT is more important to dispersion stabilization than favorable hydrophilic interactions with water. The supernatants of the LSZ−SWNT dispersions show the characteristic van Hove singularities that are suggestive of individual dispersion, as shown in Figure 4a. Previous investigations further confirmed individual dispersion using AFM and ellipsometry.28 Comparing aqueous LSZ solutions and LSZ−SWNT dispersions of equal LSZ concentrations show changes in the LSZ spectral profile in the presence of SWNT. As apparent from the black color of the dispersions, SWNT-dependent increases in absorbance at all wavelengths were significant and could overshadow changes in the LSZ spectrum. Therefore, difference spectra, shown in Figure 4b, were obtained by subtracting the absorbance spectrum of a 0.5 wt % LSZ solution without SWNT. This enabled changes in the LSZ spectrum resulting from the LSZ−SWNT interaction to be extracted. At 280 nm, the wavelength where tryptophan absorbs,43 the data show an increased absorbance of approximately 0.09 for the LSZ−SWNT supernatant and approximately 0.31 for the LSZ−SWNT mixture. In addition, a hyperchromic shift was observed, which was likely due to two phenomena. First, the partial unfolding of LSZ during sonication could have caused the shift.17 When LSZ intercalates between and partially envelops SWNT, the environment surrounding the hydrophobic core of the protein is altered. More specifically, the tryptophan chromophore in the hydro-

pristine SWNT was 1.9, but it was only 1.4 for the mixture and 1.0 for the supernatant, which also indicates doping.37 No Gband shift was observed between SWNT and LSZ−SWNT mixture, but there was a 5 cm−1 shift between the pristine SWNT and LSZ−SWNT supernatant. These results are in contrast to those of Xie et al. who found that the frequency of G′ did not shift and stated that no charge transfer is expected due to the lack of a metallic center in LSZ.30 The difference in these charge transfer results are likely due to the different starting materials and processing methods used to produce the respective dispersions. In an attempt to increase the adsorption of LSZ to SWNT through hydrogen bond interactions between hydroxyl and amine or carboxyl groups, SWNT were functionalized via acid oxidation to introduce carboxyl and hydroxyl groups and increase SWNT hydrophilicity. It is well established that the introduction of carboxyl, hydroxyl, and other hydrophilic groups as a result of oxidation significantly increases SWNT dispersibility in water.40−42 Counterintuitively, the dispersibilty of oxidized SWNT (SWNT-Ox) with LSZ was much poorer than for pristine SWNT. In fact, the SWNT-Ox supernatant was virtually clear, whereas the pristine SWNT supernatant was characteristically black. Comparing the relative changes in absorbance for the initial mixtures and supernatants elucidated the marked difference between SWNT-Ox and pristine SWNT dispersed with LSZ. The ratio of [SWNT-Ox]supernatant:[SWNTOx] mixture was 23% and the ratio [SWNT] supernatant : [SWNT]mixture was 34% at 506 nm. Assuming no significant change in the extinction coefficients for SWNT-Ox and SWNT in aqueous LSZ, the supernatant concentration of SWNT-Ox was 27.0 mg/L compared to 108.0 mg/L for pristine SWNT. The possibility that the reduced dispersibility of the SWNT-Ox was due to isoelectric effects was explored based on Nepal and Geckler’s finding that the dispersion limit of individual SWNT in aqueous LSZ was pH-dependent.24 For the SWNT used in their work, the dispersion limit was nearly constant at 89 mg/L for 1.4 < pH < 8.2 using similar centrifugation conditions of 3 h at 18 000g. Near the isoelectric point, pH = 10.2, virtually no SWNT could be dispersed. However, SWNT dispersibility increased again above the isoelectric point. The SWNT mixtures and supernatants in the present work had very similar 10345

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348

The Journal of Physical Chemistry C

Article

around 335 nm. In addition to effects of solvent polarity on the fluorescence of tryptophan, binding can cause two changes: shifts in the emission maximum wavelength and in fluorescence intensity.47,48 Horsely et al.48 were able to determine the quenching constant for bound LSZ by probing the tryptophan excitation and showing that it increased when bound to variously charged silica surfaces. Therefore, it would be expected that the tryptophan emission fluorescence intensity would decrease, or quench, in LSZ−SWNT dispersions if the tryptophan residue was bound to SWNT. The LSZ and LSZ− SWNT supernatant were excited at 280 nm to specifically probe the tryptophan fluorescence within LSZ for proof of LSZ interaction with SWNT and specific interaction of tryptophan with SWNT. Since the absorption of excitation light at 280 nm by SWNT could result in false interpretation of fluorescence quenching, the extinction coefficients of both LSZ and SWNT were determined through differential absorbance at varying dilution factors (see Supporting Information) and used to obtain a corrected LSZ−SWNT supernatant emission spectrum. Figure 6 shows the fluorescence emission spectrum of the

phobic core is no longer buried within the LSZ core. The emergence of the chromophore from the hydrophobic core could increase the absorbance of light at 280 nm. Second, π−π stacking of an aromatic ring with the sidewall of SWNT could have changed the electronic structure of the chromophore. Consequently, the increase in absorbance intensity of the chromophore at a given wavelength could also be due to the overlap of p-orbitals excited by UV-light.44 Of the three amino acids showing Cotton effects between 250 and 350 nm in CD spectra (tryptophan, tyrosine, and phenylalanine), tryptophan is twice as abundant in LSZ and absorbs about 4-fold more light at 280 nm compared to tyrosine; phenylalanine is also less abundant and has negligible absorbance at 280 nm.45 This allows CD to be used for relatively selective examination of the interaction between tryptophan and the sidewall of SWNT. Specifically, CD was used to show that the addition of SWNT causes a definitive change in the optical polarity of LSZ. The supernatant of LSZ− SWNT dispersion was used to compare the CD spectra to enhance the spectral changes due to binding, which could be masked by excess LSZ present in the mixtures. Because of its electronic structure and chirality, the tryptophan within LSZ will naturally absorb more left polarized light than right leading to a slight increase between 280 and 300 nm wavelengths.45 As shown in Figure 5, even after sonication and centrifugation, the

Figure 6. UV−vis fluorescence spectra of 0.5 wt % LSZ solution and 0.5 wt % LSZ−0.1 wt % SWNT supernatant excited at 280 nm showing the quenching of the fluorescence intensity of the LSZ− SWNT supernatant due to the binding of excitation-probed tryptophan within LSZ with SWNT. The inset shows the results of the anisotropy measurements with error.

Figure 5. Circular dichroism spectrum of 0.5 wt % LSZ solution and 0.5 wt % LSZ−0.1 wt % SWNT supernatant, which shows the loss of the tryptophan peak between wavelengths of 280 and 300 nm due to the interaction between the LSZ tryptophan and the sidewall of SWNT. The molar ellipsticity of the LSZ solution has been upshifted to that of the LSZ−SWNT supernatant for clarity.

LSZ solution and the corrected spectrum of the LSZ−SWNT supernatant dispersion. Even after sonication, the fact that the tryptophan fluorescence within the LSZ−SWNT dispersion was still quenched by approximately 13% relative to the processed LSZ solution provides further evidence of a tryptophan−SWNT interaction. It is hypothesized that the quenching of the tryptophan fluorescence results from π−π stacking interaction, which occurs between the tryptophan residue and the sidewall of SWNT.49 To further illuminate the interaction between LSZ and SWNT, anisotropic fluorescence measurements were performed. Fluorescence anisotropy is concentration independent and ideally suited for measuring binding events, especially the association of proteins with other molecules.50 Because of changes in rotational freedom, binding of proteins to other molecules increases the dependence of their fluorescence on the polarity of the excitation light. On the basis of the concentration of LSZ and SWNT present in the sample used for anisotropy measurements, for every individual SWNT present, there were approximately 1800 LSZ molecules

aqueous LSZ solutions show a tryptophan peak between wavelengths of 280 and 300 nm; this is consistent with literature values for LSZ.46 However, the addition of SWNT quenches the optical polarity at these wavelengths by approximately 23% in the supernatant. The change in structure observed in the CD spectrum upon interaction with SWNT is believed to be due to the electronic structure and hydrophobic core environment of LSZ, specifically tryptophan, being altered as a result of the hydrophobic, π−π stacking interaction between the LSZ and SWNT. The final methods used to elucidate the specific interaction between LSZ and SWNT were fluorescence spectroscopy and anisotropy. The dominant fluorophore within LSZ is tryptophan, which has an excitation of 280 nm. Tryptophan’s maximum emission wavelength is highly sensitive to the polarity of the surrounding environment but is generally 10346

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348

The Journal of Physical Chemistry C

Article

reduced activity. Solid assemblies made by coagulating mixtures extruded out of a 21 gauge syringe needle with 1-butanol had significantly lower activities of 39.0 and 31.7%, respectively. Even these seemingly low values for the solid assemblies are greater than a commercially available mouthwash, which maintains only 28.5% of native LSZ activity.56,57 The reduction in activity for SWNT solid assemblies can be related to not all the LSZ being present on the outer accessible surface of the material. If the solid assemblies were normalized to the accessible surface bound LSZ and not to the total LSZ, the activity would be much higher. Therefore, the major difference between the dispersions and solid assemblies is likely to be surface to volume ratio and LSZ accessibility dependent.

present. Even if full coverage of each SWNT is assumed, only 400 LSZ molecules bind to each SWNT, leaving an excess of 1400 LSZ molecules in solution. Therefore, the change in anisotropy of LSZ bound to SWNT was not very pronounced due to the presence of unbound LSZ. When tracking the fluorescence emission at 345 nm, the anisotropy of LSZ solutions was observed to be 0.160 and that of the LSZ− SWNT supernatant solutions was 0.166, as shown in the inset of Figure 6. Although this difference of 0.006 seems small, the result was consistent over all tests and statistically significant. A t-test of the data collected for each solution revealed p < 0.0001 indicating that the means were distinctly different. In conjunction with the fluorescence quenching, the anisotropy increase shows a direct binding interaction. This was shown through emission quenching to be tryptophan interacting with the SWNT sidewall. While understanding the nature of the interaction is of fundamental scientific interest, the antibacterial activity of LSZ−SWNT has significant potential applications. LSZ is already commercially used in food packaging and mouthwash,51 and there is interest in using it for antibacterial surfaces.28,52,53 The combination of the antibacterial activity of LSZ with the mechanical and electrical properties of SWNT has the potential to enable robust multifunctional materials. Figure 7 shows the

■

CONCLUSIONS The molecular interaction between nanotubes and proteins is of growing interest due to proteins’ ability to disperse nanotubes and potential to add additional functionality to materials assembled from protein−nanotube dispersions. We have provided additional evidence of the close interaction between LSZ and SWNT and not only supported but refined previously reported hypotheses that the interaction is a hydrophobic one involving π−π stacking. The results of this work suggest that there is a direct molecular interaction between LSZ and SWNT through the indole moiety of tryptophan and the aromatic sidewall of SWNT. UV−vis absorption showed a hyperchromic shift due to exposure and electronic structure changes. CD confirmed a change in tryptophan surroundings through quenching of the tryptophan residue peak. In addition, fluorescence spectroscopy substantiated the binding interaction through quenching of fluorescence emission and increases in fluorescence anisotropy of tryptophan at 345 nm when LSZ was bound to SWNT. Finally, we have shown that LSZ−SWNT dispersions and solid materials retain a significant portion of LSZ’s native antibacterial activity. The results of this research shed light on the interactions between LSZ and nanotubes and revealed important information for the molecular design and manipulation of nanomaterials and biopolymers, which could yield a wide array of applications. These results provide a foundation for additional work on elucidating the interactions between pristine carbon materials and other biomolecules.

Figure 7. Activity at various processing steps of LSZ−SWNT shown as a percentage of native LSZ activity where each sample has been normalized based on the total amount of LSZ present as determined by either UV−vis (dispersions) or TGA (solid assemblies). It should be noted that the calculated error for each sample was less than that of the UV−vis; therefore, the error shown is that inherent in the UV−vis measurements.

■

ASSOCIATED CONTENT

S Supporting Information *

SWNT characterization, degree of SWNT carboxylation, LSZ− SWNT dispersion interaction by ATR-FTIR, Raman analysis of D:RBM vs D:G, and determination of LSZ and SWNT extinction coefficients with accompanying figures. This material is available free of charge via the Internet at http://pubs.acs.org.

activity of LSZ−SWNT dispersions and further processed solid assemblies relative to a centrifuged and sonicated LSZ solution; each sample has been normalized for the amount of LSZ present in the sample as determined by TGA. The initial LSZ− SWNT dispersions had over 99% of the native LSZ activity. This is greater than that observed by Ding et al. who achieved a maximum of 90% of native LSZ activity in the presence of silica nanotubes.54 Merli et al. found that covalently bonding LSZ to multiwalled nanotubes (MWNT) via EDC/NHS actually increased activity,20 but covalent functionalization has the disadvantages of reducing SWNT mechanical and electrical properties.55 The activity of the supernatant being lower than that of the mixture was possibly due to a reduction in unbound lysozyme. In addition, since some studies have shown that some SWNT may have inherent antimicrobial activity, the reduced SWNT concentration may also have contributed to the

■

AUTHOR INFORMATION

Corresponding Author

*Tel: (334)844-2060. Fax: (334)844-2063. E-mail: davisva@ auburn.edu. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We would like to acknowledge the National Science Foundation CAREER/PECASE Award CMMI 0846629 and the NSF REU Site in Micro/Nano-Structured Materials, 10347

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348

The Journal of Physical Chemistry C

Article

(32) Han, J. T.; Kim, S. Y.; Jeong, H. J.; Jeong, S. Y.; Lee, G. W. J. Phys. Chem. C 2009, 113, 16915−16920. (33) Parra-Vasquez, A. N. G.; Stepanek, I.; Davis, V. A.; Moore, V. C.; Haroz, E. H.; Shaver, J.; Hauge, R. H.; Smalley, R. E.; Pasquali, M. Macromolecules 2007, 40, 4043−4047. (34) Lin, Y.; Zhou, B.; Fernando, K. A. S.; Liu, P.; Allard, L. F.; Sun, Y. P. Macromolecules 2003, 36, 7199−7204. (35) Sun, Y. P.; Zhou, B.; Henbest, K.; Fu, K.; Huang, W.; Lin, Y.; Taylor, S.; Carroll, D. L. Chem. Phys. Lett. 2002, 351, 349−353. (36) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P.; Jones, D. E. H. Science of Fullerenes and Carbon Nanotubes; Academic Press: New York, 1996; Vol. 965. (37) Moonoosawmy, K. R.; Kruse, P. J. Am. Chem. Soc. 2008, 130, 13417−13424. (38) Poenitzsch, V. Z.; Winters, D. C.; Xie, H.; Dieckmann, G. R.; Dalton, A. B.; Musselman, I. H. J. Am. Chem. Soc. 2007, 129, 14724− 14732. (39) Rao, A.; Eklund, P.; Bandow, S.; Thess, A.; Smalley, R. E. Nature 1997, 388, 257−259. (40) Wang, Y.; Wu, J.; Wei, F. Carbon 2003, 41, 2939−2948. (41) Datsyuk, V.; Kalyva, M.; Papagelis, K.; Parthenios, J.; Tasis, D.; Siokou, A.; Kallitsis, I.; Galiotis, C. Carbon 2008, 46, 833−840. (42) Zhang, J.; Zou, H.; Qing, Q.; Yang, Y.; Li, Q.; Liu, Z.; Guo, X.; Du, Z. J. Phys. Chem. B 2003, 107, 3712−3718. (43) Edelhoch, H. Biochemistry 1967, 6, 1948−1954. (44) Karachevtsev, V.; Plokhotnichenko, A.; Karachevtsev, M.; Leontiev, V. Carbon 2010, 48, 3682−3691. (45) Tanaka, F.; Forster, L. S.; Pal, P. K.; Rupley, J. A. J. Biol. Chem. 1975, 250, 6977. (46) Holzwarth, G.; Doty, P. J. Am. Chem. Soc. 1965, 87, 218−228. (47) Lehrer, S.; Fasman, G. D. J. Biol. Chem. 1967, 242, 4644. (48) Horsley, D.; Herron, J.; Hlady, V.; Andrade, J. Langmuir 1991, 7, 218−222. (49) Blancafort, L.; González, D.; Olivucci, M.; Robb, M. A. J. Am. Chem. Soc. 2002, 124, 6398−6406. (50) Lakowicz, J. R.; Masters, B. R. J. Biomed. Opt. 2008, 13, 029901. (51) Carini, S.; Mucchetti, G.; Neviani, E. Microbiol., Aliments, Nutr. 1985, 3, 299−320. (52) Eby, D. M.; Luckarift, H. R.; Johnson, G. R. ACS Appl. Mater. Interfaces 2009, 1, 1553−1560. (53) Han, Z. D.; Fina, A. Prog. Polym. Sci. 2011, 36, 914−944. (54) Ding, H. M.; Shao, L.; Liu, R. J.; Xiao, Q. G.; Chen, J. F. J. Colloid Interface Sci. 2005, 290, 102−106. (55) Dyke, C. A.; Tour, J. M. J. Phys. Chem. A 2004, 108, 11151− 11159. (56) Collins, L.; Dawes, C. J. Dent. Res. 1987, 66, 1300−1302. (57) Hannig, C.; Hoch, J.; Becker, K.; Hannig, M.; Attin, T. Arch. Oral Biol. 2005, 50, 821−828.

Therapeutics, and Devices EEC-1063107 for funding. The assistance of Dr. Paul Cobine with fluorescence anisotropy measurements and Jiaming Hu with circular dichroism measurements is gratefully acknowledged.

■

REFERENCES

(1) Dresselhaus, M.; Dresselhaus, G.; Saito, R. Carbon 1995, 33, 883−891. (2) Flynn, C. E.; Lee, S.-W.; Peelle, B. R.; Belcher, A. M. Acta Mater. 2003, 51, 5867−5880. (3) Hirano, A.; Maeda, Y.; Akasaka, T.; Shiraki, K. Chem.Eur. J. 2009, 15, 9905−9910. (4) Asuri, P.; Bale, S. S.; Pangule, R. C.; Shah, D. A.; Kane, R. S.; Dordick, J. S. Langmuir 2007, 23, 12318−12321. (5) Tu, X.; Manohar, S.; Jagota, A.; Zheng, M. Nature 2009, 460, 250−253. (6) Lu, G.; Maragakis, P.; Kaxiras, E. Nano Lett. 2005, 5, 897−900. (7) Pedrosa, V. A.; Gnanaprakasa, T.; Balasubramanian, S.; Olsen, E. V.; Davis, V. A.; Simonian, A. L. Electrochem. Commun. 2009, 11, 1401−1404. (8) Badaire, S.; Zakri, C.; Maugey, M.; Derre, A.; Barisci, J. N.; Wallace, G.; Poulin, P. Adv. Mater. 2005, 17, 1673−1676. (9) Ao, G.; Nepal, D.; Aono, M.; Davis, V. A. ACS Nano 2011, 5, 1450−1458. (10) Daniel, S.; Rao, T. P.; Rao, K. S.; Rani, S. U.; Naidu, G. R. K.; Lee, H.-Y.; Kawai, T. Sens. Actuators, B 2007, 122, 672−682. (11) Mantha, S.; Pedrosa, V. A.; Olsen, E. V.; Davis, V. A.; Simonian, A. L. Langmuir 2010, 26, 19114−19119. (12) Sanchez-Pomales, G.; Santiago-Rodriguez, L.; Cabrera, C. R. J. Nanosci. Nanotechnol. 2009, 9, 2175−2188. (13) Pedrosa, V. A.; Paliwal, S.; Balasubramanian, S.; Nepal, D.; Davis, V. A.; Wild, J.; Simonian, A. L. Colloids Surf., B 2010, 77, 69−74. (14) Barisci, J. N.; Tahhan, M.; Wallace, G. G.; Badaire, S.; Vaugien, T.; Maugey, M.; Poulin, P. Adv. Funct. Mater. 2004, 14, 133−138. (15) Razal, J. M.; Gilmore, K. J.; Wallace, G. G. Adv. Funct. Mater. 2008, 18, 61−66. (16) Raffaini, G.; Ganazzoli, F. J. Appl. Biomater. Biomech. 2010, 8, 135−145. (17) Matsuura, K.; Saito, T.; Okazaki, T.; Ohshima, S.; Yumura, M.; Iijima, S. Chem. Phys. Lett. 2006, 429, 497−502. (18) Kuboki, Y.; Koshikawa, T.; Takita, H.; Fujisawa, R.; Lee, M. H.; Abe, S.; Akasaka, T.; Uo, M.; Watari, F.; Sammons, R. Dent. Mater. J. 2010, 29, 369−373. (19) Anglaret, E.; Righi, A.; Sauvajol, J. L.; Bernier, P.; Vigolo, B.; Poulin, P. Phys. B 2002, 323, 38−43. (20) Merli, D.; Ugonino, M.; Profumo, A.; Fagnoni, M.; Quartarone, E.; Mustarelli, P.; Visai, L.; Grandi, M. S.; Galinetto, P.; Canton, P. J. Nanosci. Nanotechnol. 2011, 11, 3100−3106. (21) Borzooeian, Z.; Safavi, A.; Sheikhi, M. H.; Aminlari, M.; Doroodmand, M. M. J. Exp. Nanosci. 2010, 5, 536−547. (22) Mantha, S.; Pedrosa, V. A.; Olsen, E. V.; Davis, V. A.; Simonian, A. L. Langmuir 2010, 26, 19114−19119. (23) Nie, H. Y.; Wang, H. F.; Cao, A. N.; Shi, Z. J.; Yang, S. T.; Yuan, Y. A.; Liu, Y. F. Nanoscale 2011, 3, 970−973. (24) Nepal, D.; Geckeler, K. E. Small 2006, 2, 406−412. (25) da Silva, M. A.; Areas, E. P. G. J. Colloid Interface Sci. 2005, 289, 394−401. (26) Caro, A.; Humblot, V.; Méthivier, C.; Minier, M.; Salmain, M.; Pradier, C.-M. J. Phys. Chem. B 2009, 113, 2101−2109. (27) Nepal, D.; Minus, M. L.; Kumar, S. Macromol. Biosci. 2011, 11, 859−863. (28) Nepal, D.; Balasubramanian, S.; Simonian, A. L.; Davis, V. A. Nano Lett. 2008, 8, 1896−1901. (29) Nepal, D.; Geckeler, K. E. Small 2007, 3, 1259−1265. (30) Xie, L. M.; Chou, S. G.; Pande, A.; Pande, J.; Zhang, J.; Dresselhaus, M. S.; Kong, J.; Liu, Z. F. J. Phys. Chem. C 2010, 114, 7717−7720. (31) Dougherty, D. A. Science 1996, 271, 163. 10348

dx.doi.org/10.1021/jp300242a | J. Phys. Chem. C 2012, 116, 10341−10348